Update README.md
Browse files
README.md
CHANGED
@@ -16,6 +16,16 @@ It isn't generic in a sense that it shouldn't be used for story telling, for exa
|
|
16 |
|
17 |
This model is trained on a private dataset. The high GSM8K score is **NOT** because of the MetaMath dataset.
|
18 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
19 |
Merge config:
|
20 |
```yaml
|
21 |
slices:
|
@@ -34,14 +44,4 @@ parameters:
|
|
34 |
value: [1, 0.5, 0.7, 0.3, 0]
|
35 |
- value: 0.5 # fallback for rest of tensors
|
36 |
dtype: bfloat16
|
37 |
-
```
|
38 |
-
|
39 |
-
# Prompt Format:
|
40 |
-
|
41 |
-
```
|
42 |
-
<|system|>
|
43 |
-
{system_message} </s>
|
44 |
-
<|user|>
|
45 |
-
{prompt} </s>
|
46 |
-
<|assistant|>
|
47 |
```
|
|
|
16 |
|
17 |
This model is trained on a private dataset. The high GSM8K score is **NOT** because of the MetaMath dataset.
|
18 |
|
19 |
+
# Prompt Format:
|
20 |
+
|
21 |
+
```
|
22 |
+
<|system|>
|
23 |
+
{system_message} </s>
|
24 |
+
<|user|>
|
25 |
+
{prompt} </s>
|
26 |
+
<|assistant|>
|
27 |
+
```
|
28 |
+
|
29 |
Merge config:
|
30 |
```yaml
|
31 |
slices:
|
|
|
44 |
value: [1, 0.5, 0.7, 0.3, 0]
|
45 |
- value: 0.5 # fallback for rest of tensors
|
46 |
dtype: bfloat16
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
47 |
```
|